Rules won’t stop research misconduct in social science

Recent revelations that Francesca Gino, a full professor at Harvard Business School, may have committed fraud is not the first scandal to undermine the credibility of social science and nor will it be the last. The pay-off for engaging in deception and fraud in academia is quite high, while the chance of detection remains low.

In his 2019 book “Management Studies in Crisis: Fraud, Deception and Meaningless Research, ” Dennis Tourish notes that a substantial percentage of doctoral students and faculty admit to engaging in some form of questionable research practices, if not outright fraud, apparently without much concern for being caught. Indeed, Stuart Ritchie, author of the 2020 book Science Fictions: Exposing Fraud, Bias, Negligence and Hype in Science, maintains that the problem of fraud and questionable scientific practices might be widespread enough to undermine a substantial number of knowledge claims produced by social scientists. John Ioannidis, professor of medicine/health research and policy at Stanford University, went so far as to claim that most published research findings are false.

In response, many journals have imposed measures such as asking authors to make their data available and sign a declaration of responsibility for ensuring no co-author has committed fraud. Some journals have hired statistical experts specifically to identify bad behaviour. In addition, groups of researchers and universities are promoting open science practices, including mandating sharing data or requiring researchers to request permission from institutional review boards before embarking on their scientific inquiries.

But even if we could iron out all the practical problems around these remedies (such as sharing confidential data in the context of sensitive industries and monitoring co-authors), we still won’t catch all instances of misconduct. If rules, regulations and contracts could ensure prudent human behaviour, there would be no jails.

Nor can we rely on stirring people’s consciences. Individuals might plausibly reason that one instance of bad behaviour is hardly likely to undermine an entire field – and, besides, their own behaviour is not self-serving but is aimed at advancing knowledge.

The problem, of course, is that if everyone thinks and acts that way, whole fields will be undermined – especially if fraudulent claims get high levels of attention and citation, as they often do, garnering fame and senior positions for their authors.

In other words, social science faces a tragedy of the commons problem. In such scenarios, every individual in a group is better off pursing their narrow self-interest, but the group as a whole would be better off if they did not.

Pollution is another example. Everyone would be better off if people stopped driving their cars, but it is in no one’s narrow self-interest to do so. Typically, such situations are mitigated by some sort of an formal institutional response (such as legal restrictions on petrol-fuelled cars) but even on the rare occasions when human ingenuity can’t circumvent the restrictions, the informal institution of not following the formal institution may emerge. For example, drinking and riding a bicycle is illegal in the UK and the Netherlands but rarely, if ever, are drunken cyclists punished.

Yet drinking and driving a car, by contrast, has become a social taboo and is rarely undertaken even when the probability of getting caught is very small. The question is, can we make research fraud similarly distasteful to scientists by a combination of formal and informal institutions?

First, on the formal side, universities should punish researchers severely for fraud, including retracting tenure. While this might sound extreme, tenure was meant to protect the voice of faculty, allowing them to advance the frontiers of science by engaging in high-risk and unpopular research. This protection should not extend to lying.

Second, journals should relax the methodological straitjacket. A significant number of questionable practices involve fishing for statistical correlations between variables and presenting them as if these were hypothesised. Presented in this way, these claims are mistakenly considered to be more reliable than they really are, according to standard research design standards. Relaxing those (often informally imposed) standards will encourage greater transparency over what the researcher actually hypothesised.

Third, journals should more readily retract studies that are demonstrably unreliable, and they should consider blacklisting their authors. Journals should also establish a pathway for authors to report honest mistakes and issue corrections.

Less formally, engaging in fraud and questionable research practices should be stigmatised, and those caught should face strict social sanctions. This is easier said than done, however. Social sanctions have to emerge from the community, as they did for smoking and drink-driving. This will require faculty to inculcate curiosity and a genuine appreciation for rigorous science among students pursuing advanced degrees.

The practice of hiring PhD students on a project basis should therefore be reconsidered. It discourages free thinking and original data-gathering, especially when, as in some cases, data is already available, Moreover, the student has no freedom to change their mind if they find that the area of research does not interest them. Appropriate safeguards should be in place to ensure that students can switch to a different research topic.

Perhaps, combined with others, such measures will eventually reduce the substantial number of doctoral students and faculty willing to engage in research misconduct to a trickle, reinforcing the credibility of social science’s knowledge claims. Failing that, the tragedy of the commons could result in social science ultimately devolving into pseudo-science.

Akhil Bhardwaj is a senior lecturer in the School of Management at the University of Bath.